Ministral 8B
Mistral AI
- Released
- October 16, 2024
- Data date
- September 8, 2026
- Successor
- Ministral 3 8B
- Price
- Historical rates under Costs
- Input
- Text
- Output
- Text
Ministral 8B was Mistral’s compact model class for edge and local applications. The dense checkpoint offers about 8 billion parameters, 128,000 tokens of context, grouped-query attention, and attention optimized for uneven input lengths. Mistral named local translation, robotics, and private assistants; its Research License is not Apache 2.0.
Mistral’s current overview lists the line as deprecated and points to Ministral 3 8B. The released weights remain under the Mistral Research License.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | Open weightsSource |
| Access | LocalSource |
| Parameters | 8BSource |
| Context window | 128,000 tokensSource |
| License | Mistral Research LicenseSource |
| Release | 2024-10-16 (Inventory lists October 15; primary release post is used.)Source |
| Parameters | 8.0198 BSource |
| Architecture | dense transformer, 36 layers, GQA with 8 KV heads, ragged attention (128k,32k,32k,32k)Source |
| Context window | 131,072 tokensSource |
| Input | TextSource |
| Output | TextSource |
| Languages | 10Source |
| Tool calling | YesSource |
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
Arena Hard (Show measurement, test conditions, and source)
- Source value
- 70.9
- Score
- 70.9
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct table; GPT-4o judge
WildBench (Show measurement, test conditions, and source)
- Source value
- 41.3
- Score
- 41.3
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct table; GPT-4o judge
MBPP pass@1 (Show measurement, test conditions, and source)
- Source value
- 70
- Score
- 70
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct code/math table; pass@1
HumanEval pass@1 (Show measurement, test conditions, and source)
- Source value
- 76.8
- Score
- 76.8
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct code/math table; pass@1
Math maj@1 (Show measurement, test conditions, and source)
- Source value
- 54.5
- Score
- 54.5
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct code/math table; maj@1
Function calling internal bench (Show measurement, test conditions, and source)
- Source value
- 31.6
- Score
- 31.6
- Metric
- percent
- Unit
- %
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct table; Internal benchmark; exact harness not published
MTBench (Show measurement, test conditions, and source)
- Source value
- 8.3
- Score
- 8.3
- Metric
- rating
- Unit
- rating
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Ministral 8B model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Instruct table; GPT-4o judge
Popularity
16.6 / 100
Predecessor and successor
Sources and data date
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved September 8, 2026. |
| Output | Documented modalities. Retrieved September 8, 2026. |
| Research status | Research date September 8, 2026. 3 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | knowledge-cutoff Original wording No exact cutoff in checked card/release post. |
| Not found in the inspected sources | training-tokens Original wording No training-token count found. |
| Not found in the inspected sources | max-output-tokens Original wording Card examples use max_tokens but do not declare a model maximum. |
| Unresolved | AA benchmark rows Original wording No exact Artificial Analysis model page located in the checked public sitemap. |
| Additional source | Context window data |
| Additional source | API pricing |
| Additional source | Model metadata (retrieved September 8, 2026) |
| Additional source | Ministral announcement · Editorial description reviewed October 2, 2026 |
| Additional source | Mistral model overview · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |